CMK version: Checkmk Raw Edition 2.5.0-2025.07.22 (Docker)
OS version: Linux d6041cf463f6 6.8.0-88-generic #89-Ubuntu SMP PREEMPT_DYNAMIC Sat Oct 11 01:02:46 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
Error message: 1 pending change
Hi everyone,
I have a strange issue with Checkmk and the WATO “pending changes” indicator.
Environment
Checkmk (OMD site name: cmk)
Monitoring core: Nagios
Site is running fine, no load / disk issues (df -h OK)
Problem
I changed the Parameters for HTML email (notification parameters).
The change is saved, but the yellow badge with “1 pending change” in the GUI never goes away, even after clicking “Activate pending changes” multiple times.
What I have checked / tried so far
Configuration deploy works fine
OMD[cmk]:~$ cmk --debug -R
Generating configuration for core (type nagios)…
Precompiling host checks…OK
Validating Nagios configuration…OK
Restarting monitoring core…OK
Permissions and disk
OMD[cmk]:~$ ls -ld “$OMD_ROOT/etc/check_mk” “$OMD_ROOT/var/check_mk/wato”
drwxr-x— 8 cmk cmk 4096 Dec 10 12:15 /omd/sites/cmk/etc/check_mk/
drwxrwx— 7 cmk cmk 4096 Dec 10 12:26 /omd/sites/cmk/var/check_mk/wato/
Disk usage:
/dev/mapper/ubuntu–vg-ubuntu–lv 47G 30G 15G 68% /
Config files really get updated when saving
After changing the HTML email settings and saving (without activating yet):
find etc/check_mk/conf.d/wato -type f -mmin -5 -ls
→ etc/check_mk/conf.d/wato/notification_parameter.mk (timestamp updates)
So WATO does write the change into notification_parameter.mk.
Tried to “reset” pending changes manually
As site user cmk:
cd “$OMD_ROOT”
delete all pending logs in the site
find var -name ‘pending.log’ -delete
clear activation jobs
rm -rf tmp/check_mk/wato/activation/*
Then as root:
omd restart cmk
Browser cache cleared / new session.
The yellow “1” for pending changes is still there.
So:
Config changes are written to disk.
cmk --debug -R runs cleanly.
Permissions and disk space look fine.
Pending log(s) and activation queue have been removed and the site restarted.
Yet the GUI always shows 1 pending change.
Has anyone seen this behaviour before or knows what else to check (e.g. specific WATO / web logs, known bug in a certain version, …)?
Thanks ![]()